Importance of Extreme Learning Machine in the field of Query classification: A novel approach
Shashank Gugnani, Tushar Bihany, Rajendra Kumar Roul · 2014
The expandable and dynamic web which is a huge repository for information is growing at lightning speed and hence it is hard to find the relevant information from the web. Efficient algorithms reduce the burden of search engines up to a great extent. Query classification is one such aspect and thus a valuable asset for a search engine. Everyday millions of web queries are posted on the web. The main aim of the query classification is to classify web users' queries into a set of predefined categories. Classifying users' queries greatly reduces the number of documents to be searched and hence is a vibrant area of research. In this paper, we propose a new technique to classify queries using Extreme Learning Machines (ELM). ELM is becoming increasingly popular among researchers owing to its fast training speed and ease of implementation. We evaluate our technique on three large datasets and compare with other relevant machine learning algorithms. Results show that proposed technique works well for classifying the queries which demonstrate the accuracy and efficiency of our system.